Saying No to a Client Without Losing Them
How to decline a client request, a free pitch, a scope-creep ask, without the relationship souring. Three documented approaches from Maister, Enns, and Voss, plus prompts you can run today.
"Can you just do this one extra thing" rarely sounds like a crisis in the moment. It sounds like a reasonable favor, a small ask, a chance to be helpful. The problem shows up later: the scope has quietly grown, the margin has quietly shrunk, and the client who got used to free extras now treats the next one as owed rather than given. Saying no, clearly and early, is the harder and cheaper version of this conversation.
Why does the generic advice about saying no fail here?
Most advice on saying no to clients is generic assertiveness coaching: "use 'I' statements," "be firm but polite," "practice saying no in the mirror." None of it addresses the actual hard part, which is that the request usually isn't obviously unreasonable in isolation. A single extra revision, one more call, a quick favor for a long-standing client, each looks fine on its own. The skill isn't being assertive. It's recognizing the pattern before the tenth reasonable-looking ask has quietly doubled your workload for the same fee.
Generic AI prompting has the same blind spot. Ask an AI assistant to help you draft a polite decline and it will produce something diplomatic and vague, "I appreciate the request, but unfortunately I'm not able to accommodate it at this time", that doesn't actually name why, doesn't protect the relationship, and gives the client nothing to work with for next time. It sounds confident. It solves nothing.
The deeper issue is that a generic decline treats every refusal as the same move: a polite wall. In practice a no to a long-standing client who's never asked for anything unreasonable before is a different act than a no to a new client testing how much free work they can extract before the contract is even signed. Generic advice can't tell the two apart because it was never built from real cases, it was built to sound reasonable in any case, which is a different and much lower bar.
What documented approaches actually handle this?
David Maister: check your own motive before you speak
Maister's Trust Equation, developed jointly with Charles Green and Robert Galford in The Trusted Advisor (2000), models trustworthiness as Credibility plus Reliability plus Intimacy, divided by Self-Orientation. The denominator is the useful part here: a no driven by self-orientation, protecting your time, your margin, your convenience, reads completely differently to a client than a no driven by their actual interest.
Applied: Before declining a request, run the equation on yourself. If a client asks for a change that would blow the deadline for everyone else on the project, a self-oriented no sounds like "I don't have time for this." A client-oriented no sounds like "saying yes here means the work I already promised you ships late, and I don't think that trade is worth it for you." Same decision, opposite effect on trust.
The equation also explains why some technically polite declines still damage a relationship: a client can tell the difference between "I can't fit this in" and "fitting this in would cost you more than it costs me," even when the words are nearly identical. The first centers your constraint. The second centers their outcome. Maister's framework gives you a concrete way to check which one you've actually written before you hit send.
Blair Enns: some requests should be refused on principle, not judged case by case
Enns's documented position in Win Without Pitching (2010) is that certain categories of request, free strategic work, speculative pitching, open-ended scope without a corresponding price, should be refused as a matter of policy, not negotiated fresh every time. The case-by-case approach wears you down and erodes positioning one small yes at a time.
Applied: Before a client asks for a free "quick strategic recommendation" ahead of a paid engagement, you already know the answer: free strategic thinking is the entire product, and giving it away for free undercuts the paid version. The no is faster and less personal when it's a standing policy rather than a decision made under pressure in the moment.
This is also why Enns's approach and Maister's approach aren't in tension, they operate at different decision points. Maister's equation helps you phrase a no you've already decided to give. Enns's policy approach helps you decide, in advance, which requests never reach the point of needing a case-by-case decision at all. Having the policy set before the ask arrives removes most of the emotional weight from the moment itself.
Chris Voss: how you phrase the no determines whether it damages the relationship
Voss's tactical empathy and calibrated questions, documented in Never Split the Difference (2016), reframe a refusal as a question that hands the problem back to the other side rather than a flat statement that invites pushback. A "how" or "what" question keeps the other party engaged in solving the problem instead of defending their original ask.
Applied: Instead of "No, I can't take on unpaid research for this pitch," try "How am I supposed to do unpaid research and still deliver what I've already committed to the other clients on my calendar this month?" The client, not you, now has to reckon with the actual constraint.
Voss's method works because a direct no gives the other person exactly one move: push back or accept it. A calibrated question gives them a problem to solve, and solving it usually means they arrive at your boundary themselves rather than having it imposed on them. The result reads less like a wall and more like a shared constraint you're both now navigating.
How do these three approaches compare?
| Approach | From | When it fits | Watch out for |
|---|---|---|---|
| Trust Equation self-check | Maister, Green & Galford, The Trusted Advisor | The relationship matters long-term and the no needs to protect trust, not just your time | Reads as therapy-speak if you skip the concrete, specific no at the end |
| Refusal of free speculative work | Blair Enns, Win Without Pitching | The ask is scope creep or free strategic work disguised as a favor | Only works if you're genuinely willing to lose the client over it |
| Calibrated "how" questions | Chris Voss, Never Split the Difference | You want the other side to feel ownership of solving the problem your no creates | Slower than a direct no, and weak under genuine time pressure |
What can you run today?
"I'm about to say no to this client request: [describe the request]. Before I respond, check whether my reason for declining is about protecting their actual outcome or about protecting my own time and margin. Write the no two different ways, one from each motive, and show me the difference in how they'd land."
Why it works: forcing the two versions side by side makes a self-oriented no obvious before you send it, which is the moment it's cheapest to fix.
"I need to decline this request: [describe the ask]. Instead of a direct no, give me two or three "how" or "what" questions that hand the problem of this request back to the other person without me stating a flat refusal."
Why it works: a question keeps the other side in problem-solving mode instead of pushing them into defending the original ask against your rejection of it.
"Here is the original agreement or scope: [paste it]. Here is the new request: [describe it]. Tell me precisely which part of the original scope this falls outside of, so I can decline by pointing to the agreement rather than arguing from how I feel about the request."
Why it works: a specific line in a document is much harder to argue with than a general sense that something feels like too much.
What's the honest limit here?
Run all three and you'll have a sharper, less damaging no than instinct alone produces. What none of this replaces is knowing whether this particular client is actually worth the relationship cost of a hard line, that depends on referral value, history, and cash flow specifics no framework has visibility into. The frameworks improve how you decline. They don't decide who's worth declining to.
Maister's Trust Equation sits in the Consultant & Advisor category alongside Enns, David C. Baker, and Alan Weiss, the broader bench for pricing and positioning decisions like this one. For the underlying judgment, which clients are actually worth this kind of protection, Client Risk is a $149 tool built to run that call across your whole client book.
All the copy-paste prompts from this guide, and the rest of the Situations series, live at /prompts too, free to copy without reading the guide first.
Frequently asked questions
Isn't saying no to a client just bad for business?
Not if the alternative is saying yes to something that quietly costs you margin, scope, or credibility. A client who gets a fast, confident no to an unreasonable ask generally respects it more than a reluctant yes that shows up late or resentful. The business risk isn't the no itself, it's a no that's vague, defensive, or delivered without an alternative.
What if the client threatens to leave over the no?
That's useful information, not a reason to reverse the decision. A client who will only stay if you absorb unpaid scope or free strategic work was costing you before the threat; the threat just makes the cost visible. Decide with real numbers, referral value, margin, relationship history, not with the fear of the threat itself.
Can AI tell me whether this specific client is worth the no?
No. It can help you structure the tradeoff, lay out what you know about margin, hassle, and referral value, but it has no independent knowledge of this client beyond what you tell it. The judgment about whether this relationship is worth protecting through a no has to be yours.
Does the Trust Equation work if the client doesn't know what it is?
Yes. The client never needs to see the framework; it's a diagnostic for you, not a script to recite. You use it to check your own motive before you speak, then deliver the no in plain language.
What's the actual difference between Enns's approach and Voss's approach here?
Enns tells you which asks to refuse outright, free strategic work, speculative pitching, because saying yes undermines your positioning regardless of how the no is delivered. Voss is about delivery: how to phrase a no, often as a question, so the other side stays engaged instead of defensive. Use Enns to decide, Voss to deliver.
What if I genuinely can't afford to lose this client?
Then the no probably isn't a flat refusal, it's a renegotiation: a scoped yes with the price, timeline, or terms changed to match what's actually being asked. All three frameworks below still apply; the content of your no just shifts from 'no' to 'yes, under these specific terms.'
Written by Gareth Hoyle. Last updated 5 October 2026. Part of the authority.md guides library.
More guides.
Delegation for Founders Who Can't Let Go
Why founders keep doing the work they should have handed off months ago, and three documented approaches, from Grove, Drucker, and Fried, for delegating without losing quality or control.
Feedback People Actually Hear
Why most feedback gets nodded at and then ignored, and three documented approaches, from Scott, Edmondson, and Rogers, for feedback that actually changes behavior.
Making Decisions You Can't Take Back
How to think clearly about a decision with no undo button, and three documented approaches, from Kahneman, Munger, and Dalio, for irreversible calls. Not financial advice.